Prompt Canvas: Turn UI screenshots into build-ready instructions

Hi Product Hunt — I’m the independent developer behind SAINO, and today I’m launching Prompt Canvas. Prompt Canvas is a Windows design-handoff workspace for one recurring problem: screenshots are easy to share, but difficult to turn into precise implementation instructions. A message such as “move this button, make the title clearer, and add more space” still leaves an AI agent or developer guessing. Which element should move? Where should it go? How much larger should it be? Which details are intentional, and what would count as finished? Prompt Canvas lets you answer those questions visually. Import a screenshot, create the intended result on the canvas, then export both the visual reference and structured instructions for the next person or AI tool. A practical screenshot-to-handoff workflow: 1. Import a PNG, JPEG, or WebP from a file, drag and drop, or the clipboard. 2. Edit the visible UI with drawing, erasing, text, arrows, circles, rectangles, highlights, blur, mosaic, crop, and image resizing. 3. Select a region of the screenshot, grab it, resize or duplicate it, and place it where the interface should actually be. Prompt Canvas repairs the original area using its surrounding colour so the result remains readable. 4. Add the reason, priority, before-and-after state, and acceptance criteria for each change. 5. Export a prompt, change JSON, annotated image, original image, shareable HTML report, or a handoff ZIP with a manifest and image checksums. Prompt Canvas is not intended to be another general-purpose image editor. Its purpose is to make design intent usable by an implementation workflow. The canvas preserves coordinates, sizes, layer order, change details, and completion criteria. Image editing, projects, autosave, saved versions, prompt generation, JSON generation, and exports run on the user’s PC. Screenshots are not automatically uploaded to SAINO or an AI provider. Blur and mosaic tools can hide private details before a handoff is created. Direct AI analysis is optional and experimental. If a user intentionally configures a supported cloud provider or a local Ollama connection, Prompt Canvas shows the destination and content before sending. The core screenshot-to-prompt workflow does not require an AI account or API key. Prompt Canvas is useful for founders, product managers, designers, developers, QA and support teams, and clients who need to communicate visual changes precisely. Prompt Canvas 1.0.0 is proprietary software for 64-bit Windows 11. It supports Japanese and English, with an English installer. It is available as a one-time purchase for USD 29.99 internationally or JPY 4,980 including tax in Japan. The current installer is not code signed, so Microsoft Defender SmartScreen may show an unrecognized-app warning. Customers receive the expected filename and SHA-256 checksum and should verify both before installation. The application source and installer remain private. I would especially appreciate feedback on one question: when you hand a visual change to an AI agent or developer, which details are still most likely to be misunderstood?